extern/pdf-to-markdown
1
0
Fork 0
mirror of https://github.com/jzillmann/pdf-to-markdown.git synced 2024-11-25 09:13:49 +01:00
Code Issues Packages Projects Releases Wiki Activity

WIP markdown formatting for code/quote

Browse Source
This commit is contained in:
Johannes Zillmann 2017-02-18 11:46:13 +01:00
parent f93d1e4aa1
commit e7ff939351
3 changed files with 29 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

25
src/javascript/models/MarkdownElements.jsx Normal file
View File

@ -0,0 +1,25 @@
import PdfBlock from './BlockPage.jsx';
export const CODE_BLOCK = "Code/Quote";
export const HEADLINE1 = "Headline 1";
export function blockToText(block: PdfBlock) {
const text = concatTextItems(block);
switch (block.type) {
case CODE_BLOCK:
return '```\n' + text + '```'
case HEADLINE1:
return '#' + text;
default:
return text;
}
}
function concatTextItems(block: PdfBlock) {
var text = '';
block.textItems.forEach(item => {
text += item.text + '\n';
});
return text;
}

3
src/javascript/models/transformations/DetectCodeBlocks.jsx
View File

@ -4,6 +4,7 @@ import ParseResult from '../ParseResult.jsx';
import PdfBlock from '../PdfBlock.jsx'; import PdfBlock from '../PdfBlock.jsx';
import TextItemCombiner from '../TextItemCombiner.jsx'; import TextItemCombiner from '../TextItemCombiner.jsx';
import { REMOVED_ANNOTATION, ADDED_ANNOTATION } from '../Annotation.jsx'; import { REMOVED_ANNOTATION, ADDED_ANNOTATION } from '../Annotation.jsx';
import { CODE_BLOCK } from '../MarkdownElements.jsx';
//Detect quotes, code etc.. which is transformed to markdown code syntax //Detect quotes, code etc.. which is transformed to markdown code syntax
export default class DetectCodeBlocks extends ToPdfBlockViewTransformation { export default class DetectCodeBlocks extends ToPdfBlockViewTransformation {
@ -56,7 +57,7 @@ export default class DetectCodeBlocks extends ToPdfBlockViewTransformation {
block.annotation = REMOVED_ANNOTATION; block.annotation = REMOVED_ANNOTATION;
newBlocks.push(block); newBlocks.push(block);
newBlocks.push(new PdfBlock({ newBlocks.push(new PdfBlock({
type: 'Code/Quote', type: CODE_BLOCK,
annotation: ADDED_ANNOTATION, annotation: ADDED_ANNOTATION,
textItems: textCombiner.combine(block.textItems) textItems: textCombiner.combine(block.textItems)
})); }));

10
src/javascript/models/transformations/ToTextBlocks.jsx
View File

@ -3,6 +3,7 @@ import Transformation from './Transformation.jsx';
import BlockPageView from '../../components/debug/BlockPageView.jsx'; import BlockPageView from '../../components/debug/BlockPageView.jsx';
import ParseResult from '../ParseResult.jsx'; import ParseResult from '../ParseResult.jsx';
import BlockPage from '../BlockPage.jsx'; import BlockPage from '../BlockPage.jsx';
import { blockToText } from '../MarkdownElements.jsx';
export default class ToTextBlocks extends Transformation { export default class ToTextBlocks extends Transformation {
@ -18,17 +19,10 @@ export default class ToTextBlocks extends Transformation {
const blocks = []; const blocks = [];
parseResult.content.forEach(page => { parseResult.content.forEach(page => {
page.blocks.forEach(block => { page.blocks.forEach(block => {
var text = '';
block.textItems.forEach(item => {
// if (item.markdownElement) {
// text = item.markdownElement.transformText(item.text);
// }
text += '\n' + item.text;
});
const category = block.type ? block.type : 'Unknown'; const category = block.type ? block.type : 'Unknown';
blocks.push({ blocks.push({
category: category, category: category,
text: text text: blockToText(block)
}); });
}); });